Tool wear monitoring is important for improving machining precision and productivity in high speed machining. To handle this problem, many methods have been developed. A newly approach which is based on the developed convolution neural network theory was proposed in this study. Convolution neural network (CNN), one of the most representative models of deep learning, has powerful feature learning, non-linear fitting, and classification ability. In this paper, firstly, the time frequency analysis of cutting force signals collected by the dynamometer was carried out. Therefore, the signals were converted into the time-frequency images. And then, they were input into the CNN for training and testing. The highest accurate recognition rate (99.40%) was obtained by adjusting the network parameters and structures. It follows that applying CNN to tool wear monitoring is a prospective approach.
Paper
Full text
A Deep Learning Approach for High Speed Machining Tool Wear Monitoring
Semantic Scholar · Engineering · 2019
Abstract
Tool wear monitoring is important for improving machining precision and productivity in high speed machining. To handle this problem, many methods have been developed. A newly approach which is based on the developed convolution neural network theory was proposed in this study. Convolution neural network (CNN), one of the most representative models of deep learning, has powerful feature learning, non-linear fitting, and classification ability. In this paper, firstly, the time frequency analysis of cutting force signals collected by the dynamometer was carried out. Therefore, the signals were converted into the time-frequency images. And then, they were input into the CNN for training and testing. The highest accurate recognition rate (99.40%) was obtained by adjusting the network parameters and structures. It follows that applying CNN to tool wear monitoring is a prospective approach.